Re: [PATCH v2 4/5] cxl: CXL Performance Monitoring Unit driver

[Jonathan Cameron]

Hi Kan,

Thanks for another quick review.

A few comments on how I responded to feedback inline, but in short
I accepted all of it and made the obvious changes.

> > +static u16 cpmu_config_get_vid(struct perf_event *event)
> > +{
> > +	return FIELD_GET(GENMASK_ULL(63, 48), event->attr.config);
> > +}
> > +  
> 
> There should be at least two places which use GENMASK_ULL(63, 48), and
> also git, mask. It's better to replace them by macros.
> 
> Ideally, we can use macros to replace all these magic numbers for the
> format and define the macros close to cpmu_format_attr[]?
> In case, we will change the bits for a format later. We will not miss
> any places.
> 

Good idea and lo and behold one of them was mismatched.
(Filter enable was described as bit 18 of config1 but bit 14 was read).

...


> > +static const struct attribute_group *cpmu_attr_groups[] = {
> > +	&cpmu_events,
> > +	&cpmu_format_group,
> > +	&cpmu_cpumask_group,
> > +	NULL
> > +};
> > +
> > +/* If counter_idx == NULL, don't try to allocate a counter. */
> > +static int cpmu_get_event_idx(struct perf_event *event, int *counter_idx, int *event_idx)
> > +{
> > +	struct cpmu_info *info = pmu_to_cpmu_info(event->pmu);
> > +	DECLARE_BITMAP(configurable_and_free, CPMU_MAX_COUNTERS);
> > +	struct cpmu_event *cpmu_ev;
> > +	u32 mask;
> > +	u16 gid, vid;
> > +	int i;
> > +
> > +	vid = cpmu_config_get_vid(event);
> > +	gid = cpmu_config_get_gid(event);
> > +	mask = cpmu_config_get_mask(event);
> > +
> > +	cpmu_ev = cpmu_find_fixed_counter_event(info, vid, gid, mask);
> > +	if (!IS_ERR(cpmu_ev)) {
> > +		if (!counter_idx)
> > +			return 0;
> > +		if (!info->hw_events[cpmu_ev->counter_idx]) {  
> 
> Use info->used_counter_bm?

I guess that's more obviously similar with the handling below and possibly
a small performance saving, so sure.

> 
> > +			*counter_idx = cpmu_ev->counter_idx;
> > +			return 0;
> > +		}
> > +		/* Fixed counter is in use, but maybe a configurable one? */
> > +	}
> > +
> > +	cpmu_ev = cpmu_find_config_counter_event(info, vid, gid, mask);
> > +	if (!IS_ERR(cpmu_ev)) {
> > +		if (!counter_idx)
> > +			return 0;
> > +
> > +		bitmap_andnot(configurable_and_free, info->conf_counter_bm,
> > +			info->used_counter_bm, CPMU_MAX_COUNTERS);
> > +
> > +		i = find_first_bit(configurable_and_free, CPMU_MAX_COUNTERS);
> > +		if (i == CPMU_MAX_COUNTERS)
> > +			return -EINVAL;
> > +
> > +		*counter_idx = i;
> > +		return 0;
> > +	}
> > +
> > +	return -EINVAL;
> > +}
> > +

...


> > +static void cpmu_event_start(struct perf_event *event, int flags)
> > +{
> > +	struct cpmu_info *info = pmu_to_cpmu_info(event->pmu);
> > +	struct hw_perf_event *hwc = &event->hw;
> > +	void __iomem *base = info->base;
> > +	u64 cfg, prev_cnt;
> > +
> > +	if (WARN_ON_ONCE(!(hwc->state & PERF_HES_STOPPED)))
> > +		return;
> > +
> > +	WARN_ON_ONCE(!(hwc->state & PERF_HES_UPTODATE));
> > +	hwc->state = 0;
> > +
> > +	/*
> > +	 * Currently only hdm filter control is implemnted, this code will
> > +	 * want generalizing when more filters are added.
> > +	 */
> > +	if (info->filter_hdm) {
> > +		if (cpmu_config1_hdm_filter_en(event))
> > +			cfg = cpmu_config2_get_hdm_decoder(event);
> > +		else
> > +			cfg = GENMASK(15, 0);
> > +		writeq(cfg, base + CPMU_FILTER_CFG_REG(hwc->idx, 0));
> > +	}
> > +
> > +	cfg = readq(base + CPMU_COUNTER_CFG_REG(hwc->idx));
> > +	cfg |= FIELD_PREP(CPMU_COUNTER_CFG_INT_ON_OVRFLW, 1);
> > +	cfg |= FIELD_PREP(CPMU_COUNTER_CFG_ENABLE, 1);
> > +	cfg |= FIELD_PREP(CPMU_COUNTER_CFG_EDGE, cpmu_config1_get_edge(event) ? 1 : 0);
> > +	cfg |= FIELD_PREP(CPMU_COUNTER_CFG_INVERT, cpmu_config1_get_invert(event) ? 1 : 0);
> > +
> > +	/* Fixed purpose counters have next two fields RO */
> > +	if (test_bit(hwc->idx, info->conf_counter_bm)) {
> > +		cfg |= FIELD_PREP(CPMU_COUNTER_CFG_EVENT_GRP_ID_IDX_MSK, hwc->event_base);
> > +		cfg |= FIELD_PREP(CPMU_COUNTER_CFG_EVENTS_MSK, cpmu_config_get_mask(event));
> > +	}
> > +	cfg &= ~CPMU_COUNTER_CFG_THRESHOLD_MSK;
> > +	/*
> > +	 * For events that generate only 1 count per clock the CXL 3.0 spec
> > +	 * states the threshold shall be set to 1 but if set to 0 it will
> > +	 * count the raw value anwyay?
> > +	 * There is no definition of what events will count multiple per cycle
> > +	 * and hence to which non 1 values of threshold can apply.
> > +	 * (CXL 3.0 8.2.7.2.1 Counter Configuration - threshold field definition)
> > +	 */
> > +	cfg |= FIELD_PREP(CPMU_COUNTER_CFG_THRESHOLD_MSK,
> > +			  cpmu_config1_get_threshold(event));
> > +	writeq(cfg, base + CPMU_COUNTER_CFG_REG(hwc->idx));
> > +
> > +	local64_set(&hwc->prev_count, 0);
> > +	writeq(0, base + CPMU_COUNTER_REG(hwc->idx));
> > +
> > +	if (flags & PERF_EF_RELOAD) {
> > +		prev_cnt = local64_read(&hwc->prev_count);
> > +		writeq(prev_cnt, base + CPMU_COUNTER_REG(hwc->idx));
> > +	}  
> 
> The hwc->prev_count is already unconditionally set to 0.

Good spot.  This driver has been evolving for too long so I'm missing
silly thing like this. I'll drop the PERF_EF_RELOAD handling as
it doesn't do anything useful and it seems to be a bit random
if uncore drivers have support or not.

> 
> > +
> > +	perf_event_update_userpage(event);
> > +}
...

> > +static void cpmu_event_stop(struct perf_event *event, int flags)
> > +{
> > +	struct cpmu_info *info = pmu_to_cpmu_info(event->pmu);
> > +	void __iomem *base = info->base;
> > +	struct hw_perf_event *hwc = &event->hw;
> > +	u64 cfg;
> > +
> > +	cpmu_read(event);
> > +	WARN_ON_ONCE(hwc->state & PERF_HES_STOPPED);
> > +	hwc->state |= PERF_HES_STOPPED;
> > +
> > +	cfg = readq(base + CPMU_COUNTER_CFG_REG(hwc->idx));
> > +	cfg &= ~(FIELD_PREP(CPMU_COUNTER_CFG_INT_ON_OVRFLW, 1) |
> > +		 FIELD_PREP(CPMU_COUNTER_CFG_ENABLE, 1));
> > +	writeq(cfg, base + CPMU_COUNTER_CFG_REG(hwc->idx));
> > +
> > +	if (hwc->state & PERF_HES_UPTODATE)
> > +		return;  
> 
> The PERF_HES_UPTODATE is used to update the event count if there is a
> request and the !(hwc->state & PERF_HES_UPTODATE). It can avoid a read
> if no PERF_HES_UPTODATE is requested.
> Since the read is invoked unconditionally above, I guess you don't need
> the check.

Dropped.  Even looking very locally we set the bit on the next line anyway
and setting an already set bit is harmless.

> 
> Thanks,
> Kan